When calculating Splits, the Split times must always be sequential and so a later time cannot be used for an earlier Split, but how the times and Splits are handled depends on the algorithm.

Version 1

This version of the algorithm always considers all possible Splits for the calculations, and the calculation method runs through the splits sequentially to find the first matching chip read to fulfil the split. 

For each Split a valid timeframe is calculated based on the minimum / maximum values defined for the split and any previous splits already calculated. 

In this method an earlier Split will always take priority, and so if a detection is missed, but a later chip read is recorded at this timing point then this may cause errors. 

Version 2

The newer Splits algorithm instead runs through all chip reads for the participant sequentially to find the first matching Split with a valid time frame. 

When a Split is matched, then all other Splits (before and after) will recalculate the valid timeframes based on the minimum / maximum values defined for the Split and the currently calculated Splits. 

In this method if a Split is already calculated then an earlier Split will not be filled in by a later chip read, however if unexpected chip reads are recorded then this may cause errors.